The teleost fish PepT1-type peptide transporters and their relationships with neutral and charged substrates

In teleosts, two PepT1-type (Slc15a1) transporters, i.e., PepT1a and PepT1b, are expressed at the intestinal level. They translocate charged di/tripeptides with different efficiency, which depends on the position of the charged amino acid in the peptide and the external pH. The relation between the position of the charged amino acid and the capability of transporting the dipeptide was investigated in the zebrafish and Atlantic salmon PepT1-type transporters. Using selected charged (at physiological pH) dipeptides: i.e., the negatively charged Asp-Gly and Gly-Asp, and the positively charged Lys-Gly and Gly-Lys and Lys-Met and Met-Lys, transport currents and kinetic parameters were collected. The neutral dipeptide Gly-Gln was used as a reference substrate. Atlantic salmon PepT1a and PepT1b transport currents were similar in the presence of Asp-Gly and Gly-Asp, while zebrafish PepT1a elicited currents strongly dependent on the position of Asp in the dipeptide and zebrafish PepT1b elicited small transport currents. For Lys- and Met-containing dipeptides smaller currents compared to Gly-Gln were observed in PepT1a-type transporters. In general, for zebrafish PepT1a the currents elicited by all tested substrates slightly increased with membrane potential and pH. For Atlantic salmon PepT1a, the transport current increased with negative potential but only in the presence of Met-containing dipeptides and in a pH-dependent way. Conversely, large currents were shown for PepT1b for all tested substrates but Gly-Lys in Atlantic salmon. This shows that in Atlantic salmon PepT1b for Lys-containing substrates the position of the charged dipeptides carrying the Lys residue defines the current amplitudes, with larger currents observed for Lys in the N-terminal position. Our results add information on the ability of PepT1 to transport charged amino acids and show species-specificity in the kinetic behavior of PepT1-type proteins. They also suggest the importance of the proximity of the substrate binding site of residues such as LysPepT1a/GlnPepT1b for recognition and specificity of the charged dipeptide and point out the role of the comparative approach that exploits the natural protein variants to understand the structure and functions of membrane transporters.


Protein preparation for docking analysis
The structures of the Homo sapiens (human) transporter PepT1 (HsPepT1) selected for this study were obtained from Protein Data Bank (Berman et al., 2003).The PDB files, identified by their PDB codes: 7PN1, 7PMX and 7PMW, represented the apoprotein (Apo) HsPepT1 in the outward facing open conformation, HsPepT1 bound to Ala-Phe in the outward facing open conformation, and the HsPepT1 bound to Ala-Phe in the outward facing occluded conformation, respectively.These structures represent the conformation of the human transporter in three subsequent moments of its transport cycle (Killer et al., 2021).
To prepare structures for docking simulations, the three Protein Data Bank files were processed through PDB fixer (online server), which is able to: i) add missing heavy atoms; ii) add missing hydrogen atoms; iii) build missing loops; iv) convert non-standard residues to their standard equivalents; v) select a single position for atoms with multiple alternate positions listed; vi) delete unwanted chains from the model; vii) delete unwanted heterogens.Afterwards, the prepared structures were uploaded in the AutoDockTools of MGLTools, a phyton-based graphical interface, useful to the preparation of the PDB format files for following Autodock Vina docking analysis.
Lastly, by means of the Autodock plugin (Seeliger and de Groot, 2010) available in PyMOL, it was possible to build the docking simulation box for each (i.e., 7PN1, 7PMX and 7PMW) protein model.

Protein-ligand complexes and molecular docking simulation
The evaluation of interactions between the HsPepT1 transporter and the selected ligands (Ala-Phe, Gly-Gln, Gly-Asp, Asp-Gly, Gly-Lys, Lys-Gly, Met-Lys, Lys-Met) via molecular docking analysis was carried out by using AutoDockVina.To start the simulations, AutoDock Vina requires that both protein targets and ligands are converted into a digital file format called "pdbqt" (Rizvi et al., 2013), which is a modified protein data bank format containing atomic charges, atom type definitions and, for ligands, topological information (rotatable bonds).Specifically, we run simultaneous multiple ligands docking for each protein by using a PERL script.In particular, we created a .conffile for Vina with an Exhaustiveness of 64 repetitions and an energy maximum range of 4 kcal/mol between the first, lower energy, pose to the tenth, higher energy, pose.All the results were collected for final evaluation of the results.

Evaluation of the results
After docking analysis, we used the software ChimeraX to generate the ligand-protein complexes of all the best ten docking score interaction for each ligand.For each complex, we generated a graphic representation of the molecular surface of the ligand in the binding pocket of the proteins and the whole ligands in a single complex.All these images are summarized in Supplemental Figure 1 (for details, see Supplemental Figure 1A).To visualize and compare the different binding energies of the ligands to the different conformations of the proteins, all docking binding affinity scores, expressed in Kcal/mol, were recorded into GraphPad Prism software, version 4.02 (GraphPad Software Inc., San Diego, CA, USA).Data are summarized in Supplemental Figure 1 (for details see, Supplemental Figure 1B).ChimeraX analysis of protein-ligand complexes as obtained by molecular docking simulation of Ala-Phe, Gly-Gln, Asp-Gly, Gly-Asp, Gly-Lys, Lys-Gly, Met-Lys and Lys-Met on the human PepT1 transporter in three structural conformations [i.e., the apoprotein (Apo) in the outward facing open conformation, 7PN1, the protein bound to the peptide in the outward facing open conformation, 7PMX, and the protein bound to the peptide in the outward facing occluded conformation, 7PMW].The figure shows the different positions of the ligands after the calculation of the lower binding energy states in the molecular docking simulation.(B) AutoDock Vina results of the binding affinity (expressed in Kcal/mol) for each predicted ligand-protein complex.Lower energy values indicate a better stability of the ligand-protein complex.

Table 1 .
Percentage of positively, negatively and/or zwitterionic microspecies present at pH 6.5 and 7.6 for each tested dipeptide (see below for details).